Based on the 10-year（2001—2010） time series of Moderate-resolution Imaging Spectroradiometer（MODIS） Normalized Difference Vegetation Index（NDVI） products and meteorological station data in the area of Southwest China

we extracted the NDVI values for the footprint of meteorological measurements and calculated the percentage of precipitation anomaly（Pa）and D index（difference between precipitation and potential evapotranspiration） as two drought indices.We then involved the information on vegetation types（Vegetation type map of China’s landmass

2000） conducted a compressive spatial-temporal regression analyses against these two meteorological drought indices and NDVI anomaly at seasonal time scales.The results showing that:（1） For most vegetation types

NDVI anomaly significantly corresponded to D index with a lag of about one month（R2>= 0.7

P<0.01）;（2） These correlations were higher for the drought-sensitive vegetation types（i.e.dry land:R2= 0.83;grassland:R2=0.71） than other types;（3） The spatial distribution of NDVI anomaly was relatively consistent with that of D index especially in drought season while it was only consistent with Pain very drought season or for drought sensitive vegetation types.